Predicting Customer Demographics in a Mobile Social Network

A network made of individuals connected based on their communication behaviour using mobile phones can be called as a Mobile Social Network. Mapping and measuring of interactions and flows between people across mobile social networks are being performed extensively in an attempt to understand the intriguing patterns of human behaviour. Such analyses can help in arriving at useful inferences for improving the accuracy of 'Targeted Advertisements'.This paper makes one such attempt to extract the demographics (i.e. age, gender and economic status) of a person based on his/her connectivity in his/her respective social network(s) and mobile phone usage over a period of time. The need for prediction arises from the fact that, for prepaid users, the demographics are either unavailable or inaccurate. The results produced are evaluated and standardized based on proven statistics pertaining to the nation considered. We use candlestick charts to compare the experimental results.